

The only thing I think it might be relevant for is coding really
It’s honestly not very good for that either. Any code it generates tends to either just not work or be really inefficient and insecure.
At best, you could use code to give a basic summary on what the code you’re looking at does and us it to spot potential bugs. But you should only treat it as a bug report where it’s a starting point on manually finding a fix. Don’t let an LLM do bugfixes on its own because it’ll most likely make several more bugs in the process














How do them boots taste, peon?